In this monthly recap, we highlight the top AI news stories that made headlines in September.
Meta’s AI Releases: From AI Stickers to Smart Glasses
Meta is making waves in AI innovation this September with a slew of releases:
AI Stickers: Meta introduces AI-powered stickers, custom-crafted for users’ chats and stories based on text prompts. They will be available for selected English-language users in WhatsApp, Messenger, Instagram, and Facebook Stories.
Image Editing with AI: New AI-powered features for Instagram — Restyle and Backdrop. Restyle lets users reimagine images by applying described visual styles, while Backdrop changes backgrounds seamlessly. Both features come with AI labels to prevent any confusion with human-generated content.
AI Studio: Meta releases its AI to people outside of the company. Meta’s open AI allows developers to craft third-party AIs for messaging services, enables businesses to create brand-aligned AIs for improved customer service, and allows creators to expand their virtual presence across Meta’s apps.
Meta AI: An advanced conversational assistant available within personal chats on WhatsApp, Messenger, and Instagram. Meta AI has access to real-time information and offers a tool for image generation.
AI Characters: Along with Meta AI, there are 28 more AI characters that have unique interests, personalities, and opinions, making interactions more engaging. Some of these AIs are created in collaboration with well-known public figures and each has profiles on Instagram and Facebook, allowing us to follow them and get to know what they’re all about.
Meta Smart Glasses: Meta is changing the game with its latest release of the Ray-Ban smart glasses. Equipped with custom-designed speakers, directional audio, and a 12 MP camera, these glasses allow for high-quality photos and videos. With Meta AI integration, the glasses support voice commands, providing a hands-free experience. Users can now effortlessly livestream to social media platforms and engage with their community.
Why it matters: All of Meta’s September releases seem to be strategic moves aimed at joining the forefront of the AI race, especially considering their previous challenges with Metaverse.
OpenAI: ChatGPT-4 With Real-time Web Browsing is out
OpenAI has unveiled four significant updates for ChatGPT, enhancing its capabilities in voice generation, image understanding, and web browsing. ChatGPT can now see and interact with image queries, hear and respond in a back-and-forth conversation, and browse the internet for real-time, up-to-date information, eliminating its previous limitation to data before September 2021.
Why it matters: OpenAI marks a significant step in expanding the capabilities of large language models.
OpenAI Introduces DALL·E 3
OpenAI releases DALL·E 3 which now understands significantly more nuances and details. DALL·E 3 is built natively on ChatGPT, making image generation user-friendly without the need for complex prompt engineering. It also helps refine prompts from short sentences to detailed paragraphs and makes image adjustments based on feedback.
Why it matters: Everypixel’s CEO Dmitry Shironosov has highlighted an intriguing aspect of this release. The absence of photorealistic images of people in the provided samples raises questions, especially given OpenAI’s claim of improved understanding of nuances and details. Is this absence coincidental? It sparks curiosity about how this AI model will tackle this issue.
Stability AI Unveils Text-to-Audio Generation
Stability AI presents a technology that allows users to create short audio clips using simple text prompts. The model was trained on over 800,000 licensed music pieces, resulting in high-quality outputs that go beyond the repetitive notes. Users can now describe the music or audio they want to hear, and the system generates it accordingly.
Why it matters: Stability AI underscores the expanding potential of AI in creative fields and offers a glimpse into the future of AI-driven music and audio production.
Microsoft’s EvoDiff: A Groundbreaking Protein-Generating AI
Microsoft has unveiled EvoDiff, an AI framework designed for protein generation. Unlike traditional design methods, EvoDiff eliminates the need for structural information about target proteins, making the whole process much faster and cost-saving. Available in open source, EvoDiff could be used to create enzymes for new therapeutics and drug delivery methods as well as new enzymes for industrial chemical reactions.
Why it matters: By open-sourcing EvoDiff, Microsoft fosters collaboration and innovation within the scientific community, ultimately benefiting humanity on a global scale.
Coke’s AI-Created Mystery Flavor
The soft-drink giant continues to leverage AI in its marketing, releasing the latest AI-created flavor, Coca-Cola Y3000. This limited-edition beverage is designed to taste like the future. Using AI, Coca-Cola not only defined the flavor but also crafted the packaging, drawing from human insights on futuristic flavors.
Why it matters: By employing AI, Coca-Cola showcases how non-tech industries can drive innovation and enhance their marketing.